[Cbe-oss-dev] [PATCH] spufs: don't yield CPU in spu_yield

Christoph Hellwig hch at lst.de
Sat Mar 24 01:12:26 EST 2007


There is no reason to yield the SPU in spu_yield - if the backing
thread reenters spu_run it gets added to the end of the runqueue for
it's priority.  So the yield is just a slowdown for the case where
we have higher priority contexts waiting.


Signed-off-by: Christoph Hellwig <hch at lst.de>

Index: linux-2.6/arch/powerpc/platforms/cell/spufs/sched.c
===================================================================
--- linux-2.6.orig/arch/powerpc/platforms/cell/spufs/sched.c	2007-03-21 17:07:27.000000000 +0100
+++ linux-2.6/arch/powerpc/platforms/cell/spufs/sched.c	2007-03-21 17:07:44.000000000 +0100
@@ -498,7 +498,6 @@ void spu_deactivate(struct spu_context *
 void spu_yield(struct spu_context *ctx)
 {
 	struct spu *spu;
-	int need_yield = 0;
 
 	if (mutex_trylock(&ctx->state_mutex)) {
 		if ((spu = ctx->spu) != NULL) {
@@ -507,13 +506,10 @@ void spu_yield(struct spu_context *ctx)
 				pr_debug("%s: yielding SPU %d NODE %d\n",
 					 __FUNCTION__, spu->number, spu->node);
 				spu_deactivate(ctx);
-				need_yield = 1;
 			}
 		}
 		mutex_unlock(&ctx->state_mutex);
 	}
-	if (unlikely(need_yield))
-		yield();
 }
 
 int __init spu_sched_init(void)



More information about the cbe-oss-dev mailing list